Tristan Barrett1, Andrew N Priest1, Edward M Lawrence1, Debra A Goldman2, Anne Y Warren3, Vincent J Gnanapragasam4,5, Evis Sala1,6, Ferdia A Gallagher1. 1. 1 Department of Radiology, Addenbrooke's Hospital and the University of Cambridge, Hills Rd, Cambridge, CB2 0QQ, UK. 2. 2 Department of Epidemiology and Biostatistics, Memorial Sloan Kettering Cancer Center, New York, NY. 3. 3 Department of Histopathology, Addenbrooke's Hospital and the University of Cambridge, Cambridge, United Kingdom. 4. 4 Department of Urology, Addenbrooke's Hospital, Cambridge, United Kingdom. 5. 5 Department of Oncology, Translational Prostate Cancer Group, University of Cambridge, Hutchinson-MRC Research Centre, Cambridge, United Kingdom. 6. 6 Present address: Department of Radiology, Memorial Sloan Kettering Cancer Center, New York, NY.
Abstract
OBJECTIVE: The purpose of this study was to investigate the ability of the apparent diffusion coefficient (ADC) ratio of tumor to normal prostate tissue to overcome inherent variability based on choice of b values, with whole-mount histopathologic analysis as the reference standard for tumor identification. MATERIALS AND METHODS: Thirty-nine patients with prostate cancer underwent 3-T MRI, including DWI with b values of 0, 150, 750, and 1000 s/mm(2). ADC maps were derived from four b value combinations. Histologically derived ROIs were defined for prostate tumor and benign prostate tissue to generate a ratio. The concordance correlation coefficient was used to evaluate agreement and reproducibility at different b values. Bland-Altman plots were used to evaluate the pattern of relative measurement difference between b value combinations. The relationship between ADC values and Gleason score was tested by Spearman rank correlation. RESULTS: ADC values varied depending on the b value combination selected. The concordance correlation coefficient was higher for ADC ratios (0.883; 95% CI, 0.816-0.927) compared with absolute ADC values for normal tissue (0.873; 95% CI, 0.799-0.921) and tumor (0.792; 95% CI, 0.688-0.864). The ADC ratio concordance correlation coefficient for transition zone tumors was considerably higher than that for the peripheral zone in all cases. Bland-Altman analysis showed higher variation for ADC maps incorporating a b value of zero for both ratio and absolute values. There was a stronger inverse relationship to Gleason score for ADC ratios (rho, -0.354 to -0.456) compared with absolute ADC values (rho, -0.117 to -0.379). CONCLUSION: The use of a simple ratio of prostate tumor ADC to normal tissue ADC improved the concordance between different b value combinations and could provide a more robust means of assessing restricted diffusion in the prostate.
OBJECTIVE: The purpose of this study was to investigate the ability of the apparent diffusion coefficient (ADC) ratio of tumor to normal prostate tissue to overcome inherent variability based on choice of b values, with whole-mount histopathologic analysis as the reference standard for tumor identification. MATERIALS AND METHODS: Thirty-nine patients with prostate cancer underwent 3-T MRI, including DWI with b values of 0, 150, 750, and 1000 s/mm(2). ADC maps were derived from four b value combinations. Histologically derived ROIs were defined for prostate tumor and benign prostate tissue to generate a ratio. The concordance correlation coefficient was used to evaluate agreement and reproducibility at different b values. Bland-Altman plots were used to evaluate the pattern of relative measurement difference between b value combinations. The relationship between ADC values and Gleason score was tested by Spearman rank correlation. RESULTS: ADC values varied depending on the b value combination selected. The concordance correlation coefficient was higher for ADC ratios (0.883; 95% CI, 0.816-0.927) compared with absolute ADC values for normal tissue (0.873; 95% CI, 0.799-0.921) and tumor (0.792; 95% CI, 0.688-0.864). The ADC ratio concordance correlation coefficient for transition zone tumors was considerably higher than that for the peripheral zone in all cases. Bland-Altman analysis showed higher variation for ADC maps incorporating a b value of zero for both ratio and absolute values. There was a stronger inverse relationship to Gleason score for ADC ratios (rho, -0.354 to -0.456) compared with absolute ADC values (rho, -0.117 to -0.379). CONCLUSION: The use of a simple ratio of prostate tumor ADC to normal tissue ADC improved the concordance between different b value combinations and could provide a more robust means of assessing restricted diffusion in the prostate.
Entities:
Keywords:
DWI; Gleason score; MRI; prostate cancer
Authors: Sonia Gaur; Stephanie Harmon; Lauren Rosenblum; Matthew D Greer; Sherif Mehralivand; Mehmet Coskun; Maria J Merino; Bradford J Wood; Joanna H Shih; Peter A Pinto; Peter L Choyke; Baris Turkbey Journal: AJR Am J Roentgenol Date: 2018-05-07 Impact factor: 3.959
Authors: Tan B Nguyen; Alexander Ushinsky; Albert Yang; Michael Nguyentat; Sara Fardin; Edward Uchio; Chandana Lall; Thomas Lee; Roozbeh Houshyar Journal: Br J Radiol Date: 2018-06-21 Impact factor: 3.039
Authors: Jorge Abreu-Gomez; Daniel Walker; Tareq Alotaibi; Matthew D F McInnes; Trevor A Flood; Nicola Schieda Journal: Eur Radiol Date: 2020-03-24 Impact factor: 5.315
Authors: Tristan Barrett; Mary McLean; Andrew N Priest; Edward M Lawrence; Andrew J Patterson; Brendan C Koo; Ilse Patterson; Anne Y Warren; Andrew Doble; Vincent J Gnanapragasam; Christof Kastner; Ferdia A Gallagher Journal: Eur Radiol Date: 2017-12-08 Impact factor: 5.315